Renewable and Sustainable Energy Reviews 21 (2013) 659–673

Contents lists available at SciVerse ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Modeling phase change materials embedded in building enclosure: A review

Saleh Nasser AL-Saadi, Zhiqiang (John) Zhai n
Civil, Environmental, and Architectural Engineering, University of Colorado at Boulder, Boulder, CO 80309-0428, USA

a r t i c l e i n f o abstract

Article history: Thermal energy storage (TES) has the capability to absorb, store and release heat based on dynamic
Received 25 October 2012 surrounding environmental conditions. Sensible energy storage captures or releases heat with changes
Accepted 20 January 2013 in material’s temperature while latent heat is associated with a phase change at an isotherm or near
Available online 21 February 2013
isothermal temperature. Latent heat storage such as using a phase change material (PCM) gains
Keywords: growing attentions recently due to its ability of storing significant thermal energy within a small
Thermal energy storage volume, making it one of most promising technologies for developing energy efficient buildings.
Phase change material To quantify their technical and economic feasibility for building’s applications, computational models
Building enclosure of TES that can be integrated into whole building energy simulations are highly demanded. This paper
Numerical models
reviews the different modeling methods generally used for PCM simulations. A few numerical modeling
methods are observed in literature for modeling PCMs including the enthalpy method, the heat capacity
method, the temperature transforming model, and the heat source method. The study compares and
highlights the advantages, disadvantages and limitations of these models and methods. It particularly
explores the viability of these methods for building applications. The paper further reviews the PCM
models that have been integrated into prevalent whole building simulation programs such as
EnergyPlus, TRNSYS, ESP-r etc. The study reveals that the heat capacity method is mostly used in
programs, despite of its limitations on time and spatial resolutions. Further research is found necessary
to identify the efficiency and accuracy of these methods in building applications.
& 2013 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
2. General formulation of phase change problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
3. Numerical formulation of phase change problems using fixed grid methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660
3.1. The enthalpy method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 661
3.2. The heat capacity method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
3.2.1. The analytical/empirical relationships. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
3.2.2. The numerical approximations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
3.3. The temperature transforming model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
3.4. The heat source method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
3.5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
4. Models for building enclosures with PCM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
4.1. The simplified models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664
4.2. The intermediate models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664
4.2.1. The enthalpy method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 664
4.2.2. The heat capacity method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
4.2.3. The heat source method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
4.3. The sophisticated models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666
4.4. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666
5. PCM models integrated into whole building simulation programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 667
5.1. EnergyPlus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 667

n
Corresponding author. Tel.: þ1 303 4924699; fax: þ1 303 4927317.
E-mail address: [email protected] (Z.(Zhai).

1364-0321/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.rser.2013.01.024
660 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

5.2. TRNSYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668

5.3. ESP-r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
5.4. BSim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 668
5.5. Other building simulation programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
5.5.1. RADCOOL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
5.5.2. ESim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
5.5.3. PCMExpress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
5.6. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670

1. Introduction must be met. For pure materials there is a clear distinction

between the solid and liquid phase separated by a sharp moving
Thermal energy storage (TES) or thermal mass is a property of interface and hence melting occurs at isothermal temperature.
materials that describes its ability to absorb, store and release heat For conduction-dominated heat transfer, the governing equation
depending on the surrounding environmental conditions. Traditional can be written for the solid and liquid phase, respectively, which
architecture, for example, is distinguished with its heavy weight and have to be satisfied by the Stefan condition as follows [22]:
thermally massive construction elements to moderate the indoor Heat transfer in the solid phase:
environment extremes experienced in hot or cold days. The thermal  
@T @ @T s
properties of construction elements have significantly improved r  cs  s ¼ ks  ð1Þ
@t @x @x
thermal comfort by manipulating the indoor air temperature without
the need of mechanical air conditioning systems [1]. On the other Heat transfer in the liquid phase:
hand, light weight buildings are characterised by its lower thermal  
@T @ @T
mass and thus expose to significant temperature swings, demanding r  cl  l ¼ kl  l ð2Þ
@t @x @x
high cooling and heating energy. A dynamic thermal mass such as
phase change materials (PCM) has been considered as a promising The Stefan condition that enforces the heat balance at the
technology to reduce the inherited climatic deficiency in light weight solid–liquid interface is:
buildings.    
@ @T s @ @T l
The apparent advantage of using PCMs lies on the amount of ks   n kl  n¼rLvn ð3Þ
@x @x @x @x
latent heat a thin PCM layer can store compared to that in a
sensible heat storage material such as concrete. For instance, a Very few analytical solutions are available in the closed form
wall of 25 mm thickness with PCM could store an equivalent for phase change problems and can be found in advanced heat
amount of thermal energy as a 420 mm thick concrete wall [2]. transfer books such as those by Crank [23], Alexiades and
As a result, the use of PCMs has recently attracted great attentions Solomon [24], and Özıs)ık [25]. Therefore, approximate numerical
for improving thermal and energy performance of buildings [3–8]. solutions are usually used to handle this class of problems. The
Recent studies that review PCMs in building applications can be numerical methods for addressing these problems have been
found in [9–18]. Various challenges are, however, arisen when reviewed in literature [26–29] and can be generally divided into:
using PCM in buildings, including the variety of materials avail-
able, material liability and safety, cost and economic feasibility,
design configurations, integration with other sustainable energy 1. Fixed grid method (i.e., weak solution): These methods consist
technologies, impact on thermal and energy performance. The of fixed space grids where the boundary is tracked by the use
problem can then be considered as an optimization dilemma of an auxiliary function. Different approaches are employed to
where all these counterparts’ challenges are becoming critical in account for latent heat evolution [27,29,30]. This class of
the design process. As a result, computational modeling is often methods has been widely used and therefore will be the focus
used as an effective tool to quantitatively understand and help of this paper.
resolving this optimization problem. 2. Deforming grid method or front tracking scheme (i.e., classical
This review paper adds to the already existing studies that solution or strong numerical solution): These methods allow
discuss the mathematical modeling of phase change materials for the grid nodes move along with the moving boundary layer
building applications such as those described in [19–21]. The and thus the space girds deform as the solution develops.
objective of this study is to systematically review the general Here the interface is explicitly tracked using the Stefan condi-
modeling theories and techniques for PCM, with an emphasis on tion [24].
the specific models used for simulating the thermal and energy 3. Hybrid method: These methods utilize the features of both
performance of PCMs embedded in building enclosures. Further- fixed and deforming grids which uses a fixed background grid
more, this paper reviews and summarizes the capabilities, and employs local front tracking schemes to follow the move-
limitations and validations of prevalent whole building simula- ment of the boundary [26].
tion programs that have been used for modeling phase change
materials.

3. Numerical formulation of phase change problems using

2. General formulation of phase change problems fixed grid methods

The main feature of phase change problems (i.e., Stefan An intuitive approach in solving phase change problems is to
problems) is the moving boundary where the Stefan condition explicitly follow the moving boundary using the front-tracking
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 661

Nomenclature A half range melting temperature

r density
a matrix coefficient
c specific heat capacity Subscripts and superscripts
f fluid fraction of PCM
h enthalpy A Apparent
k thermal conductivity avg average
L latent heat of fusion e east node
n the unit normal on the phase interface eff effective
S source term l liquid state
T temperature m melting
t time n iteration level
v velocity of the interface p point node
x space distance s solid state
w west node
Greek letters

methods. However, this method needs to make a priori assump- leads to the following discretized equation:
tion that the boundary is smooth or monotonic during the period nþ1 n
hp ¼ hp þ anwþ 1  T nwþ 1 þ anp þ 1  T np þ 1 þane þ 1  T ne þ 1 ð5Þ
[31]. This assumption is not always true and therefore reformu-
nþ1
lating phase change problems using the fixed grid techniques According to Eq. (5), it is clear that the current enthalpy (hp )
becomes an obvious alternative [23,30,32,33]. The Stefan condi- is dependent on the current value of temperature (T np þ 1 ) and
tion Eq.(3) within the fixed grid method is implicitly treated by therefore the enthalpy term is nonlinear. The equation cannot be
the reformulated governing equation and hence the position of solved without using proper numerical techniques to handle this
the moving boundary is known when the solution is converged. nonlinearity. This has to be solved either by nonlinear solvers
The fixed grid method is simple compared to the others, most such as the Newton’s method or by linearizing the nonlinear
versatile, convenient, adaptable and easily-programmable [24]. terms and utilizing iterative methods as fully explained by
The latent heat evolution is accounted for in the governing [24,30,31,34,35,52–56]. If a non-linear solver is selected, an
equation by using either enthalpy method [34–38], heat capacity auxiliary temperature–enthalpy function is required for Eq. (5)
method [39–42], temperature transforming model [43–46], heat and can be written for materials that change phases at specific
source method [38,47–50], or other methods [27,29,51]. The temperature range as follows [22]:
following sections will describe the widely used methods. 8h
> p
> Cs , 
> hp r C s  ðT m  A Þ
>
< hp þ Cl Cs þ L ðT m  A Þ
>
Tp ¼ 2C Cs2 A  , C l  ðT m  A Þ o hp o C s  ðT m þ A Þ þ L
3.1. The enthalpy method > l þ 2L A
>
> 2
>
>
: hp ðC s C l ÞT m L , hp Z C l  ðT m þ A Þ þ L
Cl
In the Enthalpy method, the latent and specific heat are com-
bined into an enthalpy term in the governing equation. The enthalpy ð6Þ
method was proposed by Eyres [38] to deal with variations of Alexiades and Solomon [24] have outlined numerical schemes
thermal properties with respect to temperature. For conduction- for solving phase change problems with the enthalpy method
dominated heat transfer, the governing Eqs. (1)–(3) can be reformu- using both linear and nonlinear approaches. Knoll on the other
lated into one equation where the latent heat is absorbed into the hand reviewed various approaches utilizing nonlinear solvers to
enthalpy term as follows: resolve the Stefan problem [55]. He, in particular, developed an
 
@h @ @T algorithm to solve the Stefan problem using the Jacobian-free
r ¼ k ð4Þ Newton–Krylove method and applied for two scenarios: (1) pure
@t @x @x
materials where melting occurs at isothermal temperature and
To demonstrate this method, a fully implicit control volume (2) non-isothermal case where phase change occurs at a range of
approximation of Eq. (4) for a typical grid shown in Fig. 1 melting temperature.
An alternative approach to solving the discretized Eq. (5) is
nþ1
to linearize the nonlinear term, hp ðTÞ, using the methods
explained by Patankar [57]. The discretized nonlinear equation
becomes linear with one primary dependent variable ‘‘Temperature’’
that can be iteratively solved with enthalpy using common linear
solvers such as direct methods (e.g., Gauss elimination or tri-
diagonal algorithm) or iterative methods (e.g., Gauss–Seidel
method). Shamsunder [34], for example, proposed a Gauss–Seidel
iterative scheme where the solution sweeps from west to east to
determine the state of phase change and subsequently determine
the new nodal enthalpy. The nodal temperatures are then deter-
mined based on the discrete form of the enthalpy–temperature
relationship. To avoid excessive iterations, the scheme was later
improved by introducing an over-relaxation parameter that is used
at nodes where no phase change occurs [58]. The scheme was
Fig. 1. A typical control volume grid. however intended for phase change that occurs at isothermal
662 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

temperature. An iterative Newton linearization scheme was intro-

duced by Furzeland [31]. The solution process is the same as that of
Shamsunder except that the over-relaxation parameter can be
applied at all nodes.
Iterative methods such as Gauss–Seidel are inherently slow
and computationally inefficient. Therefore, fast numerical
schemes have been introduced to improve the computational
efficiency [35,42,48]. Pham [42] proposed a method that com-
bines the features of the enthalpy and heat capacity methods. The
method consists of two steps: a prediction step followed by a
correction step as shown in Fig. 2. Based on guessed values, the
new nodal temperatures are predicted (point (2) on the graph).
The enthalpy is determined based on the predicted temperature
values. The predicted temperatures are subsequently corrected to
be consistent with the enthalpy–temperature curve (point (3) on
the graph). This temperature correction step is the key of this
method. This method is later known to be the ‘‘Quasi-Enthalpy’’
method [59].
Voller pointed out that this method might not conserve energy
Fig. 3. Corrective iterative scheme in the Enthalpy method at a node during one
at every time step [22] and a better conservative iterative scheme time step.
was proposed by Swaminathan and Voller as illustrated in Fig. 3
[35]. The method iterates the predicted and corrected intermedi-
ate values until the convergence is achieved. The method has apparent heat capacity can be written as:
been recently investigated as an alternative to overcome the  
@T @ @T
limitations of the PCM simulation algorithm implemented in r  C A ðTÞ  ¼ k ð7Þ
@t @x @x
ESP-r [60].
The method receives the popularity because the temperature
is the only prime variable that needs to be solved in the
3.2. The heat capacity method discretized form. The key in this approach lies in the heat capacity
approximation. Two methods are commonly used to approximate
The heat capacity term in the governing equation imitates the the apparent heat capacity term in Eq. (7): the analytical/empiri-
effect of enthalpy (sensible and latent heat) by increasing the heat cal relationships and the numerical approximations.
capacity value during the phase changing stage. Two approaches
are generally used to account for the latent heat liberation: the 3.2.1. The analytical/empirical relationships
apparent heat capacity [22,27,29] and the effective heat capacity The heat capacity of a PCM can be determined from the testing
[61,62]. Although the two approaches differ in the heat capacity data with differential scanning calorimeters (DSC). Manufacturers
approximation, recent literatures, however, use the terminologies of PCMs normally provide limited data pertained to their products
interchangeably. More details on the effective heat capacity such as melting temperature, heat of fusion and heat capacity at
concept are explained by Poirier [63]. solid and liquid states. Such minimal data can be used to
The apparent heat capacity method was introduced by approximate the heat capacity of a PCM using a simple direct
Hashemi and Sliepcevich [64] to solve a one-dimensional heat relationship with an introduction of fictitious melting tempera-
transfer with phase change in a mushy region. The conduction- ture range (2  A) [22,43]:
dominated one-dimensional heat transfer equation using the 8
< Cs,
> T rT m  A ðSolid regionÞ
Cs þ Cl
m A o T o Tm þ A
A L
C ¼ 2 þ 2A , T ðMushy regionÞ ð8Þ
>
:C ,
l T ZT m þ A ðLiquid regionÞ

Convergence might be an issue when solving Eq. (7), if the half

phase change range (A) is set too small or the time step is too
large. There is a possible risk of missing the latent heat contribu-
tion in a large time step. Hence, DSC testing results can be used to
form an empirical expression to approximate the heat capacity.
Fang [65], for instance, proposed a mathematical expression for
the heat capacity of paraffin—based PCM obtained from DSC.
Others have suggested and used alternative forms to approximate
the heat capacity [66–69].

3.2.2. The numerical approximations

Numerical approximation is an alternative when detailed
information about PCM’s thermal behaviors is available. Many
numerical approximations have been proposed in literature
[41,70–75]. For example, Comini [70] applied a numerical tech-
nique in the finite element method where the heat capacity was
determined using a derivative of enthalpy with respect to tem-
Fig. 2. Corrective non-iterative scheme in the Quasi–Enthalpy method at a node perature. Later, Morgan [71] has improved the relationship to
during one time step. avoid the convergence problems. When using an iterative scheme,
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 663

the heat capacity can be approximated using the successive is, however, not commonly used but offers an alternative solution
temperature and enthalpy solutions. The temporal averaging when compared to the apparent heat capacity method.
proposed by Morgan [71] is illustrated in Fig. 4 and is represented
by the following equation: 3.4. The heat source method
n n1
Dh h  h Using the heat source method, the total enthalpy in the
CA ¼ ¼ n ð9Þ
DT T  T n1 governing Eq. (4) is split into the specific heat and latent heat
On the other hand, Lemmon [73] proposed an approximation where the latent heat acts as a source term [23,47]. Eq. (4) thus
based on the space average rather than the time average becomes:
 
approach. The temporal and space average approximations are, @T @ @T @f
r  C avg  ¼ k r  L  l ð12Þ
however, prone to convergence issues unless some precautions @t @x @x @t
are taken [76]. Solutions to the limitations of the apparent heat
The method was alluded by Eyres [38] in the mid-1940s.
capacity method have been proposed in literature [33,61,77–79].
In popular schemes, the phase change front is tracked by the
Voller [22] found that the apparent heat capacity approximation
evaluation of a nodal liquid fraction field which takes a value of
based on the direct relationships are more accurate than the
0 for solid, 1 for liquid, and a value in the range of [0–1] for the
Morgan approximation used for the cases he studied.
mushy region [23,49]. With this approach, the fluid fraction is
linearized and the equation can be solved iteratively with tem-
3.3. The temperature transforming model perature. The liquid fraction can be approximated using the
following auxiliary equation [49]:
The temperature transforming model was developed by Cao 8
> 0, if T r T m A
<
and Faghri [80] to overcome the time and spatial limitations in
f l ¼ ðT L T s Þ , if T m A o T o T m þ A
ðTT s Þ
ð13Þ
the heat capacity method. The model has been used by Faghri and >
:
his co-workers for many applications [44,45,81]. The method is 1, if T Z T m þ A
also called ‘‘the improved temperature-based equivalent heat When discretizing Eq. (12) with a fully implicit scheme and
capacity method’’ [82]. While the method was tested against linearizing the source term ‘‘liquid fraction’’ at the current time
several benchmark examples, it has been reported to produce step, the discretized equation becomes linear and needs to be
inconsistent results especially when mass transfer through PCM is solved for temperature in an iterative manner with the liquid
considered. Corrections were proposed to improve the accuracy fraction. Costa [50] has used this method to numerically simulate
[81,82]. The key of this method is that the energy Eq. (4) is the latent heat thermal storage.
transformed into a nonlinear Eq. (10) with a single dependent
variable ‘‘Temperature’’ [43]. 3.5. Summary
 
@T @ @T @S
r  C ef f ðTÞ  ¼ k r  ð10Þ Different mathematical models and methods have been sug-
@t @x @x @t
gested in literature to deal with phase change problems using the
This source term is represented by the following Equation [43]. fixed grids method: enthalpy, heat capacity, temperature trans-
8 forming method, and heat source method. Every method has its
>
> Cs  A , T oT m  A
<  main distinct feature for the latent heat liberation with advantages
Cs þ Cl
m A o T o Tm þ A
L
SðTÞ ¼ 2  A þ 2 , T ð11Þ and disadvantages. Table 1 summarizes these methods, and high-
>
>
: C  A þ L, T 4T þ A lights the main feature and their advantages and disadvantages. For
l m
many reasons including computational efficiency, modeling accu-
The latent heat during the phase change stage is represented racy and flexibility in selecting solution schemes, the enthalpy
by a source term in the governing equation with the heat capacity method is merited to be an attractive mathematical model over
term similar to the apparent heat capacity method. The method others for simulating phase change problems. In particular, it
becomes appealing when the corrective iterative scheme (i.e., a
fast and energy conservative approach), or non-iterative
scheme (i.e., a quick but conservative approach at low time steps)
are implemented. To further exploit these two features for large
time steps, a quick but energy conservative approach is envisioned.

4. Models for building enclosures with PCM

A few models have been developed to solve phase change

problems on the basis of the general mathematical methods
described above. A list of the models for various engineering fields
including building applications was reviewed recently by these
studies [19,20,83–86]. This section, however, provides a more con-
centrated and in-depth review on the models that are proposed and
used for simulating PCM integrated within building enclosures.
A few innovative and sustainable designs have been proposed
by integrating PCM within building construction elements. These
designs demand different levels of model complexity to evaluate
the thermal performance of such elements. The computational
Fig. 4. Apparent heat capacity approximation at a node during one time step using models are classified hereafter as the simplified, intermediate and
iterative methods. sophisticated models. Within this context, the simplified models
664 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

Table 1
Feature, advantages and disadvantages of mathematical methods used for phase change problems.

Mathematical Main feature Advantages Disadvantages Possible solution schemes References

model for
latent heat
evolution

Enthalpy Enthalpy accounts for  Fast if proper  Difficult to handle supercooling Iterative scheme with Non-linear [24,55]
method sensible and latent heat scheme is selected problems solvers (e.g., Newton’s methods)
 Deal with sharp as  The temperature at a typical grid Linearized-Enthalpy: Corrective [22,35]
well as gradual point may oscillate with time iterative scheme
phase change Quasi-Enthalpy: Non-iterative [42]
Temperature correction scheme
Heat capacity Heat capacity accounts  Intuitive since  Lack of computational efficiency Iterative Scheme (e.g., Gauss– [22,27,29,61,63,64]
method for both sensible and dealing with one  Small time step and fine grids are Seidel iterative scheme ) if a
latent heat dependent required for accuracy proper heat capacity
variable  Difficult in handling cases where approximation is selected
‘‘Temperature’’ the phase-change temperature
 Easy to program range is small
 Suitable for  Difficult to obtain convergence with
gradual this technique, and there is always
phase change a chance that the latent heat is
underestimated
 Not applicable for cases where
phase change occurs at fixed
temperature

Temperature Heat capacity and source  Deal with sharp  Not a common method and Iterative Scheme (e.g., Gauss– [43–46,80,81]
transforming term are used to account and gradual therefore not tested to evaluate the Seidel iterative scheme ) after
method for sensible and latent phase change pros and cons linearizing the source term
heat  Handle large time
step and
course grids

Heat source Latent heat is treated as a  Intuitive due to  Requires under-relaxation and Iterative Scheme (e.g., Gauss– [23,38,47,49]
method source term separating the therefore extra efforts is needed to Seidel iterative scheme ) after
latent heat from determine the optimum linearizing the source term
sensible relaxation factor
 Deal with sharp  Lack of computational efficiency
and gradual  Problems with round off errors if
phase change melting occurs over
temperature range

are rough approximations of the physics in the phase change trend of air temperature in the conditioned room. Another simplified
process but offers quick results. The intermediate models are a physical model using the R–C network method was developed and
tradeoff between the speed of the simplified models and the validated for three wall types: light, medium and heavy with shaped-
accuracy and flexibility of the sophisticated models. The sophis- stabilized phase change material [90]. The model, however, had
ticated models are created using well validated numerical to use a genetic algorithm to identify the key model parameters:
packages that offer a choice of established and optimized numer- resistances and capacitances of the wall layers to reach an optimal
ical methods. This class offers a high level of accuracy and node distribution. When the optimal parameters are identified, the
modeling flexibility but is computationally expensive. model can be used to simulate the heat transfer process in a wall
unit that has a PCM layer. Although the model is intended to be
4.1. The simplified models simple, multiple procedures are necessary for practical applications.
The model was however implemented to evaluate the energy
Detailed models for simulating PCM within building enclo- performance of an office building with shaped-stabilized phase
sures may capture more physics of heat transfer process. How- change material embedded in a wall unit [91].
ever, simplified models are sometimes preferred to provide a
quick estimation of PCM’s thermal performance. Some simplified
4.2. The intermediate models
models have been developed with this intention [87–91].
A steady-state analytical model for evaluating the benefits of PCM
A variety of intermediate models using, respectively, the
in walls and roofs has been proposed by Kaushik [87]. The model
enthalpy method, heat capacity method and heat source methods
used the heat capacity method to represent the dynamic thermal
have been developed for one, two and three dimensional cases for
storage of PCM. The model was utilized to analyze the dynamic
building enclosures.
thermal performance of a free floating building with PCM embedded
in a south wall fac- ade [88]. The result for a typical mild winter day in
New-Delhi showed that the wall with PCM outperformed that of an 4.2.1. The enthalpy method
ordinary wall. A rough model utilizing the heat capacity method was In the enthalpy method, the enthalpy may be solved by
developed to characterize the heat transfer process and subsequently nonlinear solvers with an auxiliary function (e.g., temperature–
estimate the temperature trend in a PCM mixed with gypsum plaster enthalpy relationship) or implicitly in the governing equation
board [89]. The simplified model was able to capture the overall using linearization techniques. A theoretical analysis based on the
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 665

enthalpy method was presented in a study evaluating a PCM in different purposes using the heat capacity method have been
a wallboard for solar energy storage [92]. A semi-implicit Crank– reported in literature [106–108].
Nicolson method was used for numerical discretization, which PCMs have also been integrated in transparent building
was subsequently solved using the Newton’s method. A more envelopes such as glazed windows. An explicit one dimensional
sophisticated two dimensional finite volume heat transfer model finite-difference model based on the heat capacity method was
based on the enthalpy method was developed and validated to extended to evaluate PCM performance when integrated into a
explore the behaviors of phase change materials incorporated into double glazing system [109,110]. The developed model was
building-integrated photovoltaic (BIPV) system [93,94]. The heat validated with experimental data and then subsequently utilized
equation is solved with an auxiliary temperature–enthalpy func- to evaluate the impact of PCM on heat loss and gains.
tion. The model was utilized to perform an optimization study of While models developed for building envelope are normally for
commercially available PCM products embedded into cavity-wall one dimensional geometry, two and three dimensional heat
systems with different wall-PCM configurations [95]. In addition transfer approaches have also been suggested for modeling PCM
to simulating heat transfer process, the model has the capability using the heat capacity method. In the early 1990s, a numerical
to solve the Navier–Stokes equation (i.e., the momentum and code ‘‘WALL88’’ was proposed for modeling two dimensional
mass equations). The model was expanded later to evaluate a transient thermal transport and storage of both sensible and latent
three dimensional heat transfer process with PCM [96]. It was heat [111]. The model was validated against analytical solution
found that the 3D model does not offer additional accuracy when and experimental lab results. The model was found to give an
compared to the previously validated 2D model. Another example excellent agreement with experimental results only after allowing
of a validated finite element 3D numerical model based on the the PCM to melt over a temperature range rather than at
enthalpy method has been suggested to simulate the PCM mixed isothermal temperature. A three dimensional finite-difference
with common mortars for wall plaster [97]. heat transfer model using the heat capacity method was devel-
Using the enthalpy linearization approach, a model was oped to study the thermal performance of randomly mixed PCM
recently presented as an alternative method for a PCM algorithm and laminated PCM-wallboard systems [112,113]. Although the
in ESP-r, a whole building simulation program [60,98]. The numerical model was not validated in these papers, the simulation
MATLAB simulation environment was used to develop a one results were helpful to conclude that laminated PCM-wallboard
dimensional numerical model using a corrective iterative scheme performs thermally better than the randomly PCM-wallboard. This
proposed by Swaminathan and Voller [35] based on the enthalpy model was later validated against an experiment and found a
linearization. The customized model in MATLAB uses the finite maximum of 3% deviation from the average experimental results
volume method with a Crank–Nicholson scheme to produce a fair [114]. The model was further used to evaluate the PCM applica-
comparison to ESP-r. The model has proven to be accurate and tions in the drywall in a passive solar building [115]. The results
fast when compared to the ESP-r results for a BESTEST Case 600 confirmed the conclusions from previous studies for the applica-
model configured with PCM. tion of laminated PCM-wallboard. Optimizing the PCM distribu-
tion within building envelope is the overall goal of an energy
efficient design. The heat capacity method was adopted by an in-
house software ‘‘CODYMUR’’, developed by a team from France, to
4.2.2. The heat capacity method optimize the use of a PCM wallboard for building energy use [116].
Phase change materials for building applications such as
Paraffin melt or freeze over a temperature range compared to
pure materials where phase change occurs at fixed temperature 4.2.3. The heat source method
[15–17,85]. This property makes the heat capacity method an The heat source method is an intuitive approach due to the
attractive approach to simulating PCM in building applications. separation of specific and latent heat. An explicit one dimensional
Utilizing MATLAB package, a research group has developed an finite-difference heat transfer model for a wall with PCM was
implicit one dimensional finite difference model for PCM in inner developed using the heat source method by Athienitis [117] and
wallboard, ceiling and floor with the heat capacity method [99]. validated against a full-scale outdoor test-room with PCM gyp-
The discretized equation was solved using the Gauss–Seidel sum board at interior side. The model showed a reasonable
iterative method. Although the lab experiments were limited agreement with experimental results. A heat transfer model of a
and simulation program was incomplete at that stage, the overall newly developed hybrid thermal energy storage system (HTESS)
benefits from PCM in wallboard were evident. using PCM capsules in a wall-unit was developed and validated
A semi-implicit one dimensional finite volume heat transfer for managing solar and electric energy [118]. The numerical
model for simulating PCM in a ceiling of a room using the heat model uses the heat source method similar to that proposed by
capacity method was developed and validated by Pasupathy Voller [48] where the latent heat evolution is represented by a
[100,101]. The model was solved using the tri-diagonal matrix source term in the governing equation. The fluid fraction is the
algorithm (TDMA) with very small time step. Although the overall key to track the latent heat process.
trend of indoor air temperature was captured by the model, the Phase change materials incorporated within floor systems was
numerical results were not in a good agreement with the experi- evaluated using a one dimensional finite volume heat transfer
ments due to many limitations of the model. The same model was model based on the heat source method [119]. The numerical
later used for evaluating the PCM integrated into a roof system [102]. model was validated with a benchmark analytical solution of
The heat capacity method has also been implemented in a one Stefan problem explained by Hu [29]. After optimizing the grid
dimensional numerical model to evaluate shape-stabilized phase and the time resolution, the developed model together with an
change materials embedded with a floor heating system [103]. optimization algorithm was used to perform an optimization
The specific heat capacity was used to account for the enthalpy analysis on PCM-floor designs. A two dimensional numerical
of PCM at different temperature regimes. The model gave good model was developed based on the heat source method to
agreement results when compared to experimental data. The simulate the effect of PCM in the design of a solar passive wall
model was also used for PCM evaluations under different climates [120]. The model has been verified using benchmark cases
and various system configurations [5,104,105]. A variety of documented in literature. The model was later validated using
modeling applications of PCM embedded in floor system for experimental data performed in the Lab and found to be
666 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

unsatisfactory due to the limitation of handling the super-cooling A computational fluid dynamics (CFD) simulation package
effect inherited in the tested PCM [121]. ‘‘FLUENT’’ has been utilized to evaluate a heat source method
when using user-defined functions for heating and cooling cycles
4.3. The sophisticated models of PCM rather than using one idealized function to represent both
phenomena [128]. The results of a PCM box model utilizing the
Developing a numerical model in two or three dimensional domain two functions showed very small error (in root mean square
is complicated and difficult to be generalized for different geometries, (RMS) values) when compared with a case of using an ideal
applications and physics; hence existing simulation packages such as function for phase change. Another example of using FLUENT is
COMSOL (formerly known as FEMLAB) [122], ANSYS-FLUENT [123], reported for PCM integration into a wall cavity system [129].
HEATING [124] and others are used as convenient design tools. Heat Engineering And Transfer In Nine Geometries (HEATING)
Although, these models offer high level of flexibility, they are not is a multidimensional, general-purpose heat transfer code that
fully explored for heat transfer process with phase change. has been extensively validated under ASHRAE project RP-1145
One study has used a commercial package FEMLAB (later [130]. The code can also be used to model the phase change using
COMSOL multiphysics) to develop a wall model with phase the heat capacity method. Ahmad [131] has used the program to
change materials using the enthalpy method and heat capacity study the behaviors of PCM in wallboard of a test cell. The model
method [125]. COMSOL is a finite element simulation package was validated using experimental test results and found to agree
that allows multi-physics modeling for many engineering appli- well with experiments. The PCM research program at Oak Ridge
cations. Utilizing this package, the created model was validated National Laboratory (ORNL) has used this program to study the
against experimental results. It was found that both numerical thermal behaviors of PCM in complex two and three geometries
methods give good estimation of latent heat evolution process. in building envelope [132]. Lab tests using a heat flow meter
However, the heat capacity method found to be more precise with apparatus (HFMA) have been conducted to validate the model in
the experimental results when a narrow melting temperature HEATING. HEATING has also been used as a standard benchmark
range of 2 1C was selected. COMSOL has also been used to study numerical package to validate the finite-difference algorithm
envelope systems with PCM [126]. The numerical results from used for PCM modeling in EnergyPlus [133].
COMSOL were successfully compared with another well-
established numerical model ‘‘WUFI-5’’. COMSOL is flexible in 4.4. Summary
modeling multi-physics within irregular and complex geometries.
For example, an innovative honeycomb wallboards with PCM A variety of models for different building enclosures have been
have been modeled in a 3D domain using the heat capacity developed using various simple, intermediate and sophisticated
method [127]. The simulation results showed a very good agree- approaches. Table 2 summarizes the models, application usages,
ment with the experimental results. and validations. It is obvious that very few simplified models have

Table 2
Modeling approaches for latent heat evolution in building enclosure.

Complexity Latent heat evolution’s approach Building enclosure Modeling Solution strategy Validation References
level case studied formulation

Simplified models
Heat capacity method Wall and roof Steady-state [87,88]
analytical model
Wallboard Experimental [89]
Optimum nodes for heat capacity Wall R-C Network N/A [90,91]
distribution using genetic algorithm
Intermediate models
Enthalpy method Wall FVM: 1D Newton’s method [92]
BIPV FVM: 2D & 3D Non-linear solver Experimental [93,95,96]
Wall FVM: 1D Iterative corrective Comparative [60,98]
scheme
Heat capacity method Wall FDM: 1D G-S Experimental [99]
Ceiling/roof FVM: 1D TDMA Experimental [100–102]
Floor FDM: 1D G-S Experimental [5,103–105]
Glazed-Windows FDM: 1D EM Experimental [109,110]
Wallboard FDM:2D Analytical and [111]
experimental
Wallboard FDM: 3D Experimental [112–115]
Heat source method Wall FDM: 1D EM Experimental [117]
Wall FDM: 1D TDMA Experimental [118]
Wall FVM: 1D Iterative scheme Analytical [119]
Wall FVM: 2D TDMA Analytical, comparative [120]
and experimental
Sophisticated numerical packages
COMSOL Heat capacity method and heat source Wall FEM: 2D Experimental [125]
method
Heat capacity method Wall FEM Comparative [126]
Heat capacity method Wall FEM: 3D Experimental [127]
FLUENT Heat source method Wall FVM: 3D SIMPLE algorithm Experimental [128]
Heat source method Wall FVM: 2D Experimental [129]
HEATING Heat capacity method Wall and roof FDM: 1D, 2D and Point-successive over- Experimental [131–133]
3D relaxation iteration

FVM: finite volume method, FDM: finite difference method, FEM: finite element method, G–S: Gauss–Seidel iterative method, TDMA: tridiagonal matrix algorithm,
EM: explicit time stepping marching, R–C: resistance–capacitance.
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 667

been suggested due to the complexity of approximating the heat [71]. While the previous versions of EnergyPlus had a semi-
transfer process associated with phase change. The intermediate implicit scheme for modeling PCMs, a fully implicit scheme has
models are commonly used but are developed for specific appli- been recently added to Version 7 of the program with more
cations and to investigate explicit envelope designs. Hence, they numerical flexibility [140]. For both schemes, it is however
lack flexibility in analyzing complex and advanced design alter- recommended to use a small time step for accurate results.
natives which becomes norms for selecting optimal or near Experimental validations have been conducted for this algo-
optimal designs. Sophisticated models offer flexibility in solving rithm with mixed feelings of accuracy. Castell, for example [141],
complex and multi-physics problems but are not fully explored found that EnergyPlus simulation results did not show a good
for modeling PCMs. This is partly due to the computational agreement with the experiments when phase change materials
inefficiency. They additionally demand considerable amount of were implemented in concrete blocks. The study concluded that
detailed data inputs, lengthy model setup and validations, and the simulation results did not reflect the thermal improvement of
limited access to the source codes. PCM observed in the test cells. However, the study highlighted
Generally, all models that adopt the heat capacity or heat that the weather data used in this simulation was not represen-
source methods must be, however, used with small time steps to tative of the actual weather data.
attain acceptable accuracy and therefore slow for whole year An experimental test shed with a commercial PCM product has
simulation which is typical for building’s thermal performance been used to validate EnergyPlus (Version 5) simulation results
evaluation. In addition, many existing models ignore inherited under the climatic conditions of Phoenix, Arizona [142]. It was
characteristics of some PCMs such as hysteresis or subcooling and found that the predicted energy consumptions were half during
therefore cannot be used for this particular application. winter and slightly greater for the summer months. In addition,
the time shift was observed for a very short time span during the
month of April (3 min) and October (9 min).
5. PCM models integrated into whole building simulation Under the climatic conditions of Auckland New-Zealand, an
programs experimental study using PCM in gypsum board has been com-
pared to the simulation results of EnergyPlus using both historical
Many detailed simulation programs are nowadays available to weather data and actual measured weather data [143]. Although
assist designers, researchers, manufacturing companies to imple- EnergyPlus model using actual weather data has captured the
ment new technologies and evaluate innovative ideas that overall trend of indoor air temperature but failed to accurately
improve the energy and thermal performance of buildings. predict the actual indoor air temperature from measurements.
Detailed simulation tools perform computations on an hourly or The study highlighted that due to many parameters including air
sub-hourly bases for accurate considerations of the dynamic infiltration, the simulation results might deviate from the actual
interactions between all thermal-based elements associated with measurement.
comfort and energy consumption, including building envelope, An early successful validation of the CondFD solution algo-
HVAC systems, lighting and control devices [134]. Many building rithm used for PCM modeling has been reported by Zhuang [144]
simulation tools are listed at the U.S. Department of Energy (DOE) using two envelope systems with PCM: envelope ‘‘A’’ (one layer of
web directory [135]. The twenty prevalent whole building energy PCM: melting temperature at 40 1C) and envelope ‘‘B’’ (two layers
simulation programs that are considered accurate and capable of of PCM: one melting temperature at 40 1C and another at 33 1C).
handling the dynamic behaviors of a building and its systems are The study shows that the largest relative difference in indoor
reviewed by Crawley et al. [136]. Few whole building simulation temperature is 12.41% and the least is 0.71% between the
programs can handle the thermal performance of building envel- simulation and testing results in a sequential 36 h period on
ope with phase change materials such as EnergyPlus, TRNSYS, envelope ‘‘A’’ condition. For envelope ‘‘B’’ condition, the largest
ESP-r, and BSim. In addition, some other programs with limited relative difference is 8.33% and the least is 0.33% in a sequential
capabilities are available for modeling phase change in buildings. 72 h. It was concluded that the most important factors in redu-
The following paragraphs brief and compare the conditions of cing the discrepancies between the simulation and the test results
these programs. are to use proper actual weather data as well as using proper
material thermal characteristics. Other successful validations of
5.1. EnergyPlus EnergyPlus algorithm for PCM have been conducted by Campbell
[145] and Chan [146] using published experimental data by
EnergyPlus uses the Conduction Transfer Functions (CTF) to Kuznik [147]. For both validation studies, the indoor air tempera-
approximate heat transfer in building envelope. Since the CTF ture was found to agree well with the experimental results.
method uses the historical values of heat flux in the computation, A field test house was used to study the impacts of PCMs in
Barbour [137] has studied the possibility of using this method in building envelope and consequently used to validate the CondFD
EnergyPlus to approximate the latent heat evolution in building algorithm in EnergyPlus [148]. The test house used the cellulose
envelope. The study developed multiple sets of CTFs based on the insulation mixed with 20% PCM by mass. The study reported that
temperature of phase change materials. A switching mechanism simulated daily average heat flux through walls was within 9% of
was proposed to exchange between these sets during the simula- the field measurements. In addition, simulation results for tem-
tion. The CTF-switching algorithm was found to be within 20% perature distribution through envelope compared fairly well with
accuracy for a range of conditions typically encountered in the experimental data apart from some delayed response com-
buildings. pared to the measurement. However, EnergyPlus has given
The capability of modeling PCMs has been facilitated in unreasonable results for heat fluxes and temperature distribu-
EnergyPlus program Version 2.0 released in April 2007 by adding tions in the attic floor of the experimental house.
a conduction finite difference (CondFD) solution algorithm [138]. In addition to the validations above, the EnergyPlus’s devel-
The algorithm uses a semi-implicit finite difference scheme based oper team has performed rigorous validation and verification
on the heat capacity method with an auxiliary enthalpy– studies for general heat transfer calculations as well as the
temperature dataset to account for latent heat evolution [139]. CondFD solution algorithm [133,149,150]. These validation stu-
Using this dataset, the heat capacity is approximated using a dies used analytical benchmark solutions, comparative tests with
temporal averaging approach similar to that proposed by Morgan well-established program ‘‘HEATING’’, and experimental results.
668 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

The studies concluded that versions prior Version 7 contains two A study reported a simplified approach of simulating PCM in
bugs and subsequently will be fixed in a later version. It is also walls/ceiling and floor in TRNSYS [158]. The approach is to use the
recommended to follow guidelines and bear in mind limitations existing capability of TRSNYS to simulate a standard active wall in
in using Version 7: ‘‘TYPE56’’ (i.e., building module in TRNSYS). The key in this
approach is a user input of equivalent heat transfer coefficients
introduced in each time step of the simulation that characterizes
 The time step should be shorter than 3 min. the thermal behaviors of the wall with PCM. The model does not
 For an accurate hourly thermal performance, 1/3 of the default evaluate the real heat transfer behaviors in PCM but accurate
node space should be used. enough for modeling PCM thermal behaviors. The model has been
 Hysteresis in PCM is not modeled in EnergyPlus and therefore validated under laboratory setting conditions.
inaccurate results may be produced. On the other hand, Schranzhofer et al. [159] have developed a
PCM module ‘‘TYPE241’’ where PCM was modeled as an internal
layer based on the heat source method. TRNSYS capability was
5.2. TRNSYS utilized to model other envelope layers using the transfer func-
tion method by creating dummy contact zones between the PCM
TRNSYS is a modular program where components modules layer and the remaining layers. In this type, the PCM is modeled
‘‘TYPES’’ are linked together in which output of one type can be using external code based on finite different method with other
an input to another in the model. It has been widely used for layers modeled through CTF algorithm available internally in
modeling building and its complex systems. Due to its modular- TRNSYS ‘‘TYPE56’’. One advantage of this approach is the short
ity, users can either utilize available types in the simulation computational time needed for numerical solution but the phy-
package or develop new modules and easily integrate to the sics might not be captured well because of assumptions involved
TRNSYS simulation package. Many features have been introduced in the dummy contact zones. The model however was not
in Version 16, including a graphical user interface ‘‘Simulation validated due to a lack of appropriate experimental data.
studio’’ and the possibility to call external programs such as Kuznik et al. have recently developed a new model ‘‘TYPE260’ in
MATLAB, FLUENT and many others [151]. TRNSYS utilizing the heat capacity method [160]. The model is semi-
Many models have been proposed in TRNSYS for modeling implicit since the physical properties of PCM used in the computa-
phase change heat transfer in building envelope but majority are tions are calculated from previous time step. This type has been
proprietary research modules. Ghoneim for example used a validated with two lab tests conducted by authors: one when the
modified type of the thermal storage wall ‘‘TYPE36’’ where the outdoor temperature was increased in two steps and the second
use of PCM has been tested for thermal storage in a wall system when it was a sinusoidal behavior. The heating heat capacity curve
[152,153]. The model was based on the enthalpy method and was used for the numerical modeling. For both validations, the
solved using an explicit scheme. Despite the numerical problems simulation results showed good agreement with the test results.
encountered when modeling PCM due to the smaller time step A newly developed one-dimensional heat transfer model using
required for the stability of the explicit scheme, the model was the heat capacity method was applied to a dividing wall with 16
successfully integrated into TRNSYS and validated against pub- glass bricks filled with PCM in TRNSYS [161]. The model has been
lished data for a concrete storage wall. Another explicit numerical validated and showed fair agreement with experimental results. In
scheme using the enthalpy method was developed for modeling addition, a simplified PCM module ‘‘TYPE1270’’ has been recently
the effects of integrating PCM into a solar wall [154,155]. The developed by Thermal Energy System Specialists (TESS) and added
module ‘‘TYPE58’’ was integrated into TRNSYS to explore to its commercially available individual components [162]. The
the significance of heating outside air for ventilation purposes module simulates PCM as an internal layer within an envelope
in the experimental house. system. The model is currently limited to materials that melt/freeze
Modeling PCM in TRNSYS is recently provided through at isothermal temperature and with constant specific heat at solid
‘‘TYPE204’’ by a team of researchers from Helsinki University of and liquid. In the transition state, the PCM layer temperature is
Technology [156]. The model simulates heat transfer through a constant and the model tracks the energy absorption and release.
wall in a three dimensional domain using the Crank–Nicolson The tracking methodology is similar to the heat source method and
scheme with 729 nodes. The model can indeed use a fully implicit therefore can be identified as ‘‘Quasi-Heat Source Method’’.
or explicit scheme with an appropriate selection of a parameter
that switches between different schemes. The model uses the 5.3. ESP-r
heat capacity to account for latent heat evolution in the wall.
Although this type has not been validated in its 3D form due to its ESP-r is a dynamic energy simulation tool of UK, used for
poor computational efficiency, Ahmad [157] has converted the 3D modeling thermal, visual and acoustic performance of buildings
model into a 1D module ‘‘TYPE101’’ and validated the modified [163]. With many features suitable to model advance sustainable
code. The simulation results were compared to experimental energy technologies, ESP-r has the capability to model phase
results from two test cells: one without PCM and another with change materials using two methods: the effective heat capacity
PCM. While the model without PCM works well when compared method and the additional heat source method [62,164,165]. ESP-
to the experimental results, the model with PCM overestimates r uses four models for PCM simulation, with one that accounts for
the daily peak indoor temperature in the cell. The authors sub-cooling, using special materials function. However, it is
outlined several reasons for this discrepancy including: (i) eva- necessary to use a small time step to obtain accurate results for
luation of the energy transmitted through the window, (ii) these two methods. While simulation results using ESP-r have
imprecision in the melting temperature range taken in the heat been found in literature, none showed any substantial validations
capacity definition, (iii) values of the convective heat transfer for these two algorithms in ESP-r [166–169].
coefficient between wall surfaces and ambient air and (iv)
existence of cold bridges. Out of these reasons, it was found that 5.4. BSim
correcting cold bridges by introducing extra term for resistance
improved the simulation results significantly when compared to BSim is a dynamic simulation program originated from
the experimental results. Denmark that offers an easy user graphical interface [170]. Using
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 669

the quasi-steady state in building modeling, the program models 5.5. Other building simulation programs
phase change using the heat capacity method [171]. The simula-
tion time step has to be small, too, for accurate prediction. Lab Some other building simulation programs have been devel-
test results from literature were used to validate the model on oped for specific research purposes of modeling phase change
three cases: continuous heating, continuous cooling, and heating materials, such as RADCOOL [172], ESim [173], and CoDyBa [66].
but with initial temperature below melting point of PCM. The Few have been proposed and developed to simulate simple
simulation model captures the overall trend of actual thermal building configurations, such as PCMExpress [174] or the one
behaviors of PCM but with small deviations. using Engineering Equation Solver (EES) [106]. Due to limited

Table 3
Numerical methods for latent heat evolution in building simulation programs.

Building Module Numerical Numerical Time stepping Limitations/Constrains Validation Reference

simulation identification formulation method used for scheme
latent heat
evolution

EnergyPlus CondFD FDM: 1D Heat capacity 1. Implicit  Time step o 3 min Analytical, [133,144–146,148–150]
method 2. Semi-implicit  Small grids Comparative and
 Hysteresis in PCM is not Experimental
modeled
 Phase Change at isothermal
temperature is not modeled

TRNSYS Modified FDM: 1D Enthalpy method Explicit  Low time step Limited validation [152,153]
‘‘TYPE36’’  No access to the code using
experimental
results for
concrete
‘‘TYPE58’’ FDM: 2D Enthalpy method Explicit No access to the code Experimental [154]
‘‘TYPE204’’ FDM: 3D Heat capacity Select an appropriate Computationally inefficient N/A [156]
method factor for implicit,
semi-implicit or
explicit
‘‘TYPE101’’ FDM: 1D Heat Capacity Semi-implicit  A correction factor to Experimental [157]
Method (Crank–Nicolson) account for cold bridges has
to be used for model
accuracy

TRNSYS Equivalent Variable heat  Real heat transfer physics in Experimental [158]
‘‘Active Wall’’ heat transfer source function PCM is not modeled
coefficients mimicking PCM
behavior
‘‘TYPE241’’ FDM: 1D Heat source No Published data N/A [159]
method
‘‘TYPE260’’ FDM: 1D Heat capacity Implicit  Thermal properties including Experimental [160]
method heat capacity are based on
previous time step (i.e.,
explicit scheme)

Modified FDM: 1D Heat capacity Implicit  Developed for Internal Experimental [161]
‘‘TYPE101’’ method partition wall

‘‘TYPE1270’’ Lumped Quasi-heat source  Very simplified model N/A [162]

method method  Internal layer within an
using heat envelope
balance  Based on lumped heat
balance (not a finite volume),
low accuracy
 For phase change at fixed
temperature

ESP-r SPMCMP53- FDM: 1D Heat capacity and  Low time step N/A [164,165]
SPMCMP56 heat source
method
BSim FVM: 1D Heat capacity Implicit  Low time step to avoid Experimental [171]
method instability

RADCOOL FDM: 1D Heat capacity Implicit [172]

method
ESim FDM: 1D Heat capacity Explicit  Explicit scheme requires low Experimental [179]
method time step to avoid instability

FVM: finite volume method, FDM: finite difference method.

670 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

literature available for these programs, only few will be discussed 6. Conclusions
hereafter.
Significant heat storage offered by phase change materials is
promising and favorable for developing various innovative building
5.5.1. RADCOOL energy technologies. To quantify the technical and economic feasi-
RADCOOL is a design tool for cooling and heating system bility of PCM-embedded technologies, it requires the development
developed at the Lawrence Berkeley National Laboratory of the US of proper computational models. This study reviews various numer-
[175]. The program was created using the simulation problem ical modeling approaches of phase change problems such as the
analysis and research kernel (SPARK) [176]. A one dimensional enthalpy, heat capacity, temperature-transforming and heat source
finite-difference model for a wall with PCM was added to this methods. The main features, advantages and disadvantages of each
validated thermal building simulation program [172]. The model method have been discussed. The discretized form of the heat
was then used to study the capability of a double PCM-wallboard equation with PCM can either be solved with nonlinear solvers such
to achieve thermal comfort without using mechanical cooling as Newton’s methods or via linearizing the nonlinear term and using
system under a typical climatic condition of Sunnyvale, California linear solvers such as iterative methods. For both approaches, the
[177]. numerical solutions are computationally inefficient or difficult to
reach convergence. Therefore, fast numerical schemes are suggested
such as the quasi enthalpy non-iterative scheme or the enthalpy
5.5.2. ESim
conservative iterative scheme.
ESim was developed at University of Dayton for building
Using these general mathematical methods, different compu-
energy simulation and can be downloaded from the developer
tational models have been developed to simulate PCMs in build-
website [178]. The simulation program was expanded and vali-
ing enclosures. Based on the level of complexity, models are
dated to model PCM-wallboards using an explicit finite-difference
classified into three categories: simple, intermediate and sophis-
approach [179]. A list of template files are available for use but
ticated models. Majority of these models have been validated
with limited capability to model complex buildings and systems.
using analytical solutions, comparative testing using validated
numerical models, and/or experimental results.
5.5.3. PCMExpress While many models are used to study the heat transfer in an
PCMExpress is a planning and simulation program for build- enclosure unit, a few models have been integrated into whole
ings using phase change materials [174], which was developed by building simulation programs. A variety of models are available
a German company, in collaboration with the Fraunhofer Institute and some are available with no cost to users such as EnergyPlus,
for Solar Energy (ISE) in Freiburg and partners from industry ‘‘TYPE204’’ in TRNSYS or ESP-r. These particular models, however,
[180]. The program simulates a free floating building with a have limitations on modeling PCM including the time and spatial
library data for weather and various construction materials resolutions, inability to model hysteresis, lack of validations of
including PCM and the flexibility to add new materials. It is an some models, and poor computational efficiency. These modeling
effective tool to evaluate the economic and technical feasibility of challenges add complexity to the already existing uncertainties in
PCM usage during an early design stage. The mathematical model experimental results of PCM’s thermal behaviors. Therefore,
of the heat transfer process in PCM is not available. The model has further research is needed to quantitatively explore the prediction
been tested by Castell [141] who found that the simulations performance of different models including their limitations on
deviate significantly from the experiments. As commented by accuracy, parameters sensitivity, speed, and stability for modeling
Castell, the discrepancy could be attributed to the lack of accurate PCM envelopes under different climatic and operating conditions.
infiltration model in the program. The program however has been
used to demonstrate the impact of using PCM in residential and
commercial buildings [180,181]. Acknowledgements

This study is part of an ongoing research at University of

5.6. Summary Colorado at Boulder funded by US National Science Foundation
(EFRI-1038305). The first author would also gratefully acknowl-
Whole building simulation programs play an important role edge the financial support received from Sultan Qaboos Univer-
for studying the economic and technical feasibility of PCMs. This sity, Oman, for supporting his PhD Study.
section reviews the capability of various whole building simula-
tion programs as summarized in Table 3. It is noted that most
PCM models integrated into whole building simulation programs References
are based on the heat capacity method. Hence, it becomes
necessary to reduce the typical one hour time step to a very [1] Givoni B. Man, climate and architecture. Van Nostrand Reinhold; 1981.
small time step (i.e., in order of minutes) to achieve acceptable [2] Telkes M. Thermal energy storage in salt hydrates. Solar Energy Materials
1980;2:381–93.
level of accuracy. For one year thermal performance evaluation, [3] Koschenz M, Lehmann B. Development of a thermally activated ceiling
building simulation programs become computationally inefficient panel with PCM for application in lightweight and retrofitted buildings.
since iterative methods are used in each time step. Additionally, Energy and Buildings 2004;36:567–78.
[4] Zalba B, Marı!n JM, Cabeza LF, Mehling H. Free-cooling of buildings with
the convergence may not be achieved due to numerical instability phase change materials. International Journal of Refrigeration 2004;27:
especially when PCM enters or leaves the phase change region. 839–49.
With all constraints and limitations above, none of whole building [5] Zhang YP, Lin KP, Yang R, Di HF, Jiang Y. Preparation, thermal performance
and application of shape-stabilized PCM in energy efficient buildings.
simulation programs are currently implementing efficient math-
Energy and Buildings 2006;38:1262–9.
ematical models that are quick, accurate and numerically stable [6] Arkar C, Vidrih B, Medved S. Efficiency of free cooling using latent heat
at realistic time step. It becomes important to thoroughly inves- storage integrated into the ventilation system of a low energy building.
tigate different mathematical models with various numerical International Journal of Refrigeration 2007;30:134–43.
[7] Cabeza LF, Castellón C, Nogués M, Medrano M, Leppers R, Zubillaga O. Use of
approaches for modeling PCMs in whole building simulation microencapsulated PCM in concrete walls for energy savings. Energy and
programs. Buildings 2007;39:113–9.
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 671

[8] Castell A, Martorell I, Medrano M, Pérez G, Cabeza LF. Experimental study of [40] Morgan K. A numerical analysis of freezing and melting with convection.
using PCM in brick constructive solutions for passive cooling. Energy and Computer Methods in Applied Mechanics and Engineering 1981;28:
Buildings 2010;42:534–40. 275–84.
[9] Baetens R, Jelle BP, Gustavsen A. Phase change materials for building applica- [41] Hsiao JS, Chung BTF. An efficient algorithm for finite element solution to
tions: a state-of-the-art review. Energy and Buildings 2010;42:1361–8. two-dimensional heat transfer with melting and freezing. Journal of Heat
[10] Khudhair AM, Farid MM. A review on energy conservation in building Transfer 1986;108:462–4.
applications with thermal storage by latent heat using phase change [42] Pham QT. A fast, unconditionally stable finite-difference scheme for heat
materials. Energy Conversion and Management 2004;45:263–75. conduction with phase change. International Journal of Heat and Mass
[11] Zhang Y, Zhou G, Lin K, Zhang Q, Di H. Application of latent heat thermal Transfer 1985;28:2079–84.
energy storage in buildings: state-of-the-art and outlook. Building and [43] Faghri A, Zhang Y. Transport phenomena in multiphase systems. Elsevier
Environment 2007;42:2197–209. Academic Press; 2006.
[12] Tyagi VV, Buddhi D. PCM thermal storage in buildings: a state of art. [44] Zeng X, Faghri A. Temperature-transforming model for binary solid–liquid
Renewable and Sustainable Energy Reviews 2007;11:1146–66. phase-change problems. Part I: Mathematical modeling and numerical
[13] Pasupathy A, Velraj R, Seeniraj RV. Phase change material-based building methodology. Numerical Heat Transfer Part B: Fundamentals 1994;25:
architecture for thermal management in residential and commercial estab- 467–80.
lishments. Renewable and Sustainable Energy Reviews 2008;12:39–64. [45] Zeng X, Faghri A. Temperature-transforming model for binary solid–liquid
[14] Zhu N, Ma Z, Wang S. Dynamic characteristics and energy performance of phase-change problems. Part II: Numerical simulation. Numerical Heat
buildings using phase change materials: a review. Energy Conversion and Transfer, Part B: Fundamentals 1994;25:481–500.
Management 2009;50:3169–81. [46] Cao Y, Faghri A, Soon CW. A numerical analysis of Stefan problems for
[15] Kuznik F, David D, Johannes K, Roux J-J. A review on phase change materials generalized multi-dimensional phase-change structures using the enthalpy
integrated in building walls. Renewable and Sustainable Energy Reviews transforming model. International Journal of Heat and Mass Transfer
2011;15:379–91. 1989;32:1289–98.
[16] Cabeza LF, Castell A, Barreneche C, de Gracia A, Fernández AI. Materials used [47] Voller VR. Implicit finite—difference solutions of the enthalpy formulation
as PCM in thermal energy storage in buildings: a review. Renewable and of Stefan problems. IMA Journal of Numerical Analysis 1985;5:201–14.
Sustainable Energy Reviews 2011;15:1675–95. [48] Voller VR. Fast implicit finite-difference method for the analysis of phase
[17] Zhou D, Zhao CY, Tian Y. Review on thermal energy storage with phase change problems. Numerical Heat Transfer, Part B: Fundamentals
change materials (PCMs) in building applications. Applied Energy 1990;17:155–69.
2012;92:593–605. [49] Swaminathan CR, Voller VR. Towards a general numerical scheme for
[18] Osterman E, Tyagi VV, Butala V, Rahim NA, Stritih U. Review of PCM based solidification systems. International Journal of Heat and Mass Transfer
cooling technologies for buildings. Energy and Buildings 2012;49:37–49. 1997;40:2859–68.
[19] Dutil Y, Rousse DR, Salah NB, Lassue S, Zalewski L. A review on phase- [50] Costa M, Buddhi D, Oliva A. Numerical simulation of a latent heat thermal
change materials: mathematical modeling and simulations. Renewable and energy storage system with enhanced heat conduction. Energy Conversion
Sustainable Energy Reviews 2011;15:112–30. and Management 1998;39:319–30.
[20] Agyenim F, Hewitt N, Eames P, Smyth M. A review of materials, heat [51] Salcudean M, Abdullah Z. On the numerical modelling of heat transfer
transfer and phase change problem formulation for latent heat thermal during solidification processes. International Journal for Numerical Meth-
energy storage systems (LHTESS). Renewable and Sustainable Energy ods in Engineering 1988;25:445–73.
Reviews 2010;14:615–28. [52] McAdie RL, Cross JT, Lewis RW, Gethin DT. A finite element enthalpy
[21] Sharma A, Tyagi VV, Chen CR, Buddhi D. Review on thermal energy storage technique for solving coupled nonlinear heat conduction/mass diffusion
with phase change materials and applications. Renewable and Sustainable problems with phase change. International Journal of Numerical Methods
Energy Reviews 2009;13:318–45. for Heat & Fluid Flow 1995;5:907–21.
[22] Voller VR. An overview of numerical methods for solving phase change [53] Kelley CT. Iterative methods for linear and nonlinear equations: Society for
problems. In: Minkowycz WJ, Sparrow EM, editors. Advances in numerical Industrial and Applied Mathematics; 1995.
heat transfer. Taylor & Francis; 1997. p. 341–80. [54] Williams MA, Wilson DG. Iterative solution of a nonlinear system arising in
[23] Crank J. Free and moving boundary problems. Clarendon Press; 1984. phase-change problems. SIAM Journal on Scientific Computing 1990;11:
[24] Alexiades V, Solomon AD. Mathematical modeling of melting and freezing 1087–101.
processes:. Hemisphere Pub. Corp.; 1993. [55] Knoll DA, Keyes DE. Jacobian-free Newton–Krylov methods: a survey of
[25] Özıs)ık N. Heat conduction. Wiley; 1993. approaches and applications. Journal of Computational Physics 2004;193:
[26] Voller VR, Swenson JB, Kim W, Paola C. An enthalpy method for moving 357–97.
boundary problems on the earth’s surface. International Journal of Numer- [56] Kelley CT. Solving nonlinear equations with Newton’s method. Society for
ical Methods for Heat & Fluid Flow 2006;16:641–54. Industrial and Applied Mathematics 2003.
[27] Idelsohn S, Storti M, Crivelli L. Numerical methods in phase-change [57] Patankar SV. Numerical heat transfer and fluid flow:. Hemisphere Pub.
problems. Archives of Computational Methods in Engineering 1994;1: Corp.; 1980.
49–74. [58] Shamsundar N, Rooz E. Numerical methods for moving boundary problems.
[28] Basu B, Date A. Numerical modelling of melting and solidification In: Minkowyex WJ, Sparrow EM, Schneider GE, Pletcher RH, editors. Hand-
problems—a review. Sadhana 1988;13:169–213. book of numerical heat transfer. New York: Wiley-Interscience; 1988.
[29] Hu H, Argyropoulos SA. Mathematical modelling of solidification and [59] Pham QT. Modelling heat and mass transfer in frozen foods: a review.
melting: a review. Modelling and Simulation in Materials Science and International Journal of Refrigeration 2006;29:876–88.
Engineering 1996;4:371–96. [60] Sadasivam S, Almeida F, Zhang D, Fung AS. An iterative enthalpy method to
[30] Voller VR, Swaminathan CR, Thomas BG. Fixed grid techniques for phase overcome the limitations in ESP-r’s PCM solution algorithm. ASHRAE
change problems: a review. International Journal for Numerical Methods in Transactions 2011;117:100–7.
Engineering 1990;30:875–98. [61] Poirier D, Salcudean M. On numerical methods used in mathematical model-
[31] Furzeland RMA. Comparative study of numerical methods for moving ing of phase change in liquid metals. Journal of Heat Transfer 1988;110:
boundary problems. IMA Journal of Applied Mathematics 1980;26:411–29. 562–70.
[32] Clavier L, Arquis E, Caltagirone JP, Gobin D. A fixed grid method for the [62] Heim D. Isothermal storage of solar energy in building construction.
numerical solution of phase change problems. International Journal for Renewable Energy 2010;35:788–96.
Numerical Methods in Engineering 1994;37:4247–61. [63] Poirier DJ. On numerical methods used in mathematical modelling of phase
[33] Rolph WD, Bathe K-J. An efficient algorithm for analysis of nonlinear heat change in liquid metals. Canada: University of Ottawa; 1986.
transfer with phase changes. International Journal for Numerical Methods [64] Hashemi HT, Sliepcevich CM. A numerical method for solving two-
in Engineering 1982;18:119–34. dimensional problems of heat conduction with change of phase. Chemical
[34] Shamsundar N, Sparrow EM. Analysis of multidimensional conduction Engineering ProgSympseries 1967;63:34–41.
phase change via the enthalpy model. Journal of Heat Transfer 1975;97: [65] Fang Y, Medina MA. Proposed modifications for models of heat transfer
333–40. problems involving partially melted phase change processes. Journal of
[35] Swaminathan CR, Voller VR. On the enthalpy method. International Journal ASTM International 2009:6.
of Numerical Methods for Heat & Fluid Flow 1993;3:233–44. [66] Virgone J, Noël J, Reisdorf R, Lyon U. Numerical study of the influence of the
[36] Albasiny EL. The solution of non-linear heat-conduction problems on the thickness and melting point on the effectiveness of phase change materials:
pilot ace. Proceedings of the IEE—Part B: Radio and Electronic Engineering application to the renovation of a low inertia school Building Simulation
1956;103:158–62. 2009, In: 11th international IBPSA conference. Glasgow, Scotland, 2009.
[37] White RE. An enthalpy formulation of the stefan problem. SIAM Journal on [67] Egolf PW, Manz H. Theory and modeling of phase change materials with
Numerical Analysis 1982;19:1129–57. and without mushy regions. International Journal of Heat and Mass
[38] Eyres NR, Hartree DR, Ingham J, Jackson R, Sarjant RJ, Wagstaff JB. The Transfer 1994;37:2917–24.
calculation of variable heat flow in solids. Philosophical Transactions of the [68] Feustel HE. Simplified numerical description of latent storage character-
Royal Society of London Series A, Mathematical and Physical Sciences istics for phase change wallboard. Lawrence Berkeley Laboratory Report,
1946;240:1–57. LBL-36933 University of California, Berkeley; 1995.
[39] Bonacina C, Comini G, Fasano A, Primicerio M. Numerical solution of phase- [69] Fachinotti VD, Cardona A, Huespe AE. A fast convergent and accurate
change problems. International Journal of Heat and Mass Transfer temperature model for phase-change heat conduction. International Journal
1973;16:1825–32. for Numerical Methods in Engineering 1999;44:1863–84.
672 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673

[70] Comini G, Del Guidice S, Lewis RW, Zienkiewicz OC. Finite element solution building simulation 2011: 12th conference of international building per-
of non-linear heat conduction problems with special reference to phase formance simulation association, Sydney; 2011.
change. International Journal for Numerical Methods in Engineering [99] Chen C, Guo H, Liu Y, Yue H, Wang C. A new kind of phase change material
1974;8:613–24. (PCM) for energy-storing wallboard. Energy and Buildings 2008;40:882–90.
[71] Morgan K, Lewis RW, Zienkiewicz OC. An improved algrorithm for heat [100] Pasupathy A, Velraj R. Mathematical modeling and experimental study on
conduction problems with phase change. International Journal for Numer- building ceiling system incorporating phase change material (PCM) for
ical Methods in Engineering 1978;12:1191–5. energy conservation, In: ASME conference proceedings; 2006. p. 59–68.
[72] Del Giudice S, Comini G, Lewis RW. Finite element simulation of freezing [101] Pasupathy A, Velraj R. Effect of double layer phase change material in
processes in soils. International Journal for Numerical and Analytical building roof for year round thermal management. Energy and Buildings
Methods in Geomechanics 1978;2:223–35. 2008;40:193–203.
[73] Lemmon E. Multidimensional integral phase change approximations for finite [102] Pasupathy A, Athanasius L, Velraj R, Seeniraj RV. Experimental investigation
element conduction codes. In: Lewis RW, Morgan K, Zienkiewicz OC, editors. and numerical simulation analysis on the thermal performance of a building
Numerical methods in heat transfer. New York: Wiley; 1981. p. 201–13. roof incorporating phase change material (PCM) for thermal management.
[74] Pham QT. The use of lumped capacitance in the finite-element solution of Applied Thermal Engineering 2008;28:556–65.
heat conduction problems with phase change. International Journal of Heat [103] Lin K, Zhang Y, Xu X, Di H, Yang R, Qin P. Modeling and simulation of under-
and Mass Transfer 1986;29:285–91. floor electric heating system with shape-stabilized PCM plates. Building and
[75] Comini G, Giudice SD, Saro O. A conservative algorithm for multidimen- Environment 2004;39:1427–34.
sional conduction phase change. International Journal for Numerical Meth- [104] Xu X, Zhang Y, Lin K, Di H, Yang R. Modeling and simulation on the thermal
ods in Engineering 1990;30:697–709. performance of shape-stabilized phase change material floor used in passive
[76] Pham QT. Comparison of general-purpose finite-element methods for the solar buildings. Energy and Buildings 2005;37:1084–91.
Stefan problem. Numerical Heat Transfer, Part B: Fundamentals 1995;27: [105] Zhou G, Zhang Y, Wang X, Lin K, Xiao W. An assessment of mixed type PCM-
417–35. gypsum and shape-stabilized PCM plates in a building for passive solar
[77] Yao M, Chait A. An alternative formulation of the apparent heat capacity heating. Solar Energy 2007;81:1351–60.
method for phase-change problems. Numerical Heat Transfer, Part B: [106] Mazo J, Delgado M, Marin JM, Zalba B. Modeling a radiant floor system with
Fundamentals 1993;24:279–300. phase change material (PCM) integrated into a building simulation tool:
[78] Gong Z-X, Mujumdar AS. Non-convergence versus non-conservation in analysis of a case study of a floor heating system coupled to a heat pump.
effective heat capacity methods for phase change problems. International Energy and Buildings 2012;47:458–66.
Journal of Numerical Methods for Heat & Fluid Flow 1997;7:565–79. [107] Jin X, Zhang X. Thermal analysis of a double layer phase change material
[79] Storti M, Crivelli LA, Idelsohn SR. An efficient tangent scheme for solving floor. Applied Thermal Engineering 2011;31:1576–81.
phase-change problems. Computer Methods in Applied Mechanics and [108] Ansuini R, Larghetti R, Giretti A, Lemma M. Radiant floors integrated with
Engineering 1988;66:65–86. PCM for indoor temperature control. Energy and Buildings 2011;43:
[80] Cao Y, Faghri AA. Numerical analysis of phase-change problems including 3019–26.
natural convection. Journal of Heat Transfer 1990;112:812–6. [109] Weinläder H, Pottler K, Beck A, Fricke J. Angular-dependent measurements
[81] Wang S, Faghri A, Bergman TL. A comprehensive numerical model for of the thermal radiation of the sky. High Temperatures–High Pressures
melting with natural convection. International Journal of Heat and Mass 2002;34:185–92.
Transfer 2010;53:1986–2000. [110] Weinläder H, Beck A, Fricke J. PCM-facade-panel for daylighting and room
[82] Zhanhua M, Yuwen Z. Solid velocity correction schemes for a temperature heating. Solar Energy 2005;78:177–86.
transforming model for convection phase change. International Journal of [111] Kedl RJ. Conventional wallboard with latent heat storage for passive solar
Numerical Methods for Heat & Fluid Flow 2006;16:204–25. applications. Energy conversion engineering conference 1990, In: IECEC-90
[83] Verma P, Varun, Singal SK. Review of mathematical modeling on latent heat Proceedings of the 25th intersociety; 1990. p. 222–225.
thermal energy storage systems using phase-change material. Renewable [112] Kim JS, Darkwa K. Simulation of an integrated PCM–wallboard system.
and Sustainable Energy Reviews 2008;12:999–1031. International Journal of Energy Research 2003;27:215–23.
[84] Regin AF, Solanki SC, Saini JS. Heat transfer characteristics of thermal energy [113] Kim JS, Darkwa J. Enhanced performance of laminated PCM wallboard for
storage system using PCM capsules: a review. Renewable and Sustainable thermal energy storage in buildings, In: energy conversion engineering
Energy Reviews 2008;12:2438–58. conference; 2002. p. 647–651.
[85] Zalba B, Marı!n JM, Cabeza LF, Mehling H. Review on thermal energy storage [114] Darkwa K, Kim JS. Dynamics of energy storage in phase change drywall
with phase change: materials, heat transfer analysis and applications. systems. International Journal of Energy Research 2005;29:335–43.
Applied Thermal Engineering 2003;23:251–83. [115] Darkwa K, O’Callaghan PW. Simulation of phase change drywalls in a
[86] Caldwell J, Kwan YY. Numerical methods for one-dimensional Stefan passive solar building. Applied Thermal Engineering 2006;26:853–8.
problems. Communications in Numerical Methods in Engineering 2004;20: [116] Kuznik F, Virgone J, Noel J. Optimization of a phase change material
535–45. wallboard for building use. Applied Thermal Engineering 2008;28:1291–8.
[87] Kaushik SC, Sodha MS, Bhardwaj SC, Kaushik ND. Periodic heat transfer and [117] Athienitis AK, Liu C, Hawes D, Banu D, Feldman D. Investigation of the
load levelling of heat flux through a PCCM thermal storage wall/roof in an thermal performance of a passive solar test-room with wall latent heat
air-conditioned building. Building and Environment 1981;16:99–107. storage. Building and Environment 1997;32:405–10.
[88] Chandra S, Kumar R, Kaushik S, Kaul S. Thermal performance of a non-air- [118] Ait Hammou Z, Lacroix M. A new PCM storage system for managing
conditioned building with PCCM thermal storage wall. Energy Conversion simultaneously solar and electric energy. Energy and Buildings 2006;38:
and Management 1985;25:15–20. 258–65.
[89] Voelker C, Kornadt O, Ostry M. Temperature reduction due to the applica- [119] Arnault A, Mathieu-Potvin F, Gosselin L. Internal surfaces including phase
tion of phase change materials. Energy and Buildings 2008;40:937–44. change materials for passive optimal shift of solar heat gain. International
[90] Zhu N, Wang S, Xu X, Ma Z. A simplified dynamic model of building Journal of Thermal Sciences 2010;49:2148–56.
structures integrated with shaped-stabilized phase change materials. Inter- [120] Joulin A, Younsi Z, Zalewski L, Rousse DR, Lassue S. A numerical study of the
national Journal of Thermal Sciences 2010;49:1722–31. melting of phase change material heated from a vertical wall of a
[91] Zhu N, Wang S, Ma Z, Sun Y. Energy performance and optimal control of air- rectangular enclosure. International Journal of Computational Fluid
conditioned buildings with envelopes enhanced by phase change materials. Dynamics 2009;23:553–66.
Energy Conversion and Management 2011;52:3197–205. [121] Joulin A, Younsi Z, Zalewski L, Lassue S, Rousse DR, Cavrot J-P. Experimental
[92] Drake JB. A study of the optimal transition temperature of PCM (phase and numerical investigation of a phase change material: thermal-energy
change material) wallboard for solar energy storage; 1987. storage and release. Applied Energy 2011;88:2454–62.
[93] Huang MJ, Eames PC, Norton B. Thermal regulation of building-integrated [122] COMSOL Inc. URL: /http://www.comsol.com/S [accessed: October/23/
photovoltaics using phase change materials. International Journal of Heat 2012].
and Mass Transfer 2004;47:2715–33. [123] ANSYS Inc.—FLUENT. URL: /http://www.ansys.com/S. [accessed: October/
[94] Huang MJ. The effect of using two PCMs on the thermal regulation 23/2012].
performance of BIPV systems. Solar Energy Materials and Solar Cells [124] The Radiation Safety Information Computational Center (RSICC)—HEATING.
2011;95:957–63. URL: /http://www-rsicc.ornl.gov/codes/psr/psr1/psr-199.htmlS. [accessed:
[95] Huang MJ, Eames PC, Hewitt NJ. The application of a validated numerical October/23/2012].
model to predict the energy conservation potential of using phase change [125] Lamberg P, Lehtiniemi R, Henell A-M. Numerical and experimental
materials in the fabric of a building. Solar Energy Materials and Solar Cells investigation of melting and freezing processes in phase change material
2006;90:1951–60. storage. International Journal of Thermal Sciences 2004;43:277–87.
[96] Huang MJ, Eames PC, Norton B. Comparison of predictions made using a [126] Baghban MH, Hovde PJ, Gustavsen A. Numerical simulation of a building
new 3D phase change material thermal control model with experimental envelope with high performance materials, In: COMSOL conference. Paris;
measurements and predictions made using a validated 2D model. Heat 2010.
Transfer Engineering 2007;28:31–7. [127] Hasse C, Grenet M, Bontemps A, Dendievel R, Sallée H. Realization, test and
[97] Sá AV, Azenha M, de Sousa H, Samagaio A. Thermal enhancement of modelling of honeycomb wallboards containing a phase change material.
plastering mortars with phase change materials: experimental and numer- Energy and Buildings 2011;43:232–8.
ical approach. Energy and Buildings 2012;49:16–27. [128] Gowreesunker BL, Tassou SA, Kolokotroni M. Improved simulation of phase
[98] Almeida F, Zhang D, Fung AS, Leong WH. Comparison of corrective phase change processes in applications where conduction is the dominant heat
change material algorithm with ESP-r simulation, In: Proceedings of transfer mode. Energy and Buildings 2012;47:353–9.
 S.N. Al-Saadi, Z.(. Zhai / Renewable and Sustainable Energy Reviews 21 (2013) 659–673 673

[129] Izquierdo-Barrientos MA, Belmonte JF, Rodrı́guez-Sánchez D, Molina AE, [155] Stritih U. Heat transfer enhancement in latent heat thermal storage system
Almendros-Ibáñez JA. A numerical study of external building walls containing for buildings. Energy and Buildings 2003;35:1097–104.
phase change materials (PCM). Applied Thermal Engineering 2012;47:73–85. [156] Jokisalo J, Lamberg P, Siren K. Suitability of building construction materials
[130] American society of heating refrigerating air-conditioning engineers, Ener- in short-term energy storage-office room simulations, In: Proceedings of IEA
modal Engineering Limited, Oak Ridge National Laboratory, Polish Academy annex 10-PCMs and chemical reactions for thermal energy storage, 3rd
of Sciences. Modeling two-and three-dimensional heat transfer through workshop. Finland; 1999. p. 11–18.
composite wall and roof assemblies in transient energy simulation pro- [157] Ahmad M, Bontemps A, Sallée H, Quenard D. Thermal testing and numerical
grams (1145—TRP): Part I: final Report: American Society of Heating, simulation of a prototype cell using light wallboards coupling vacuum
Refrigerating and Air-Conditioning Engineers; 2001. isolation panels and phase change material. Energy and Buildings
[131] Ahmad M, Bontemps A, Sallée H, Quenard D. Experimental investigation and 2006;38:673–81.
computer simulation of thermal behaviour of wallboards containing a phase [158] Ibáñez M, Lázaro A, Zalba B, Cabeza LF. An approach to the simulation of
change material. Energy and Buildings 2006;38:357–66. PCMs in building applications using TRNSYS. Applied Thermal Engineering
[132] Kosny J, Stovall TK, Shrestha SS, DW Yarbrough . Theoretical and experi- 2005;25:1796–807.
mental thermal performance analysis of complex thermal storage mem- [159] Schranzhofer H, Puschnig P, Heinz A, Streicher W. Validation of a TRNSYS
brane containing bio-based phase change material (PCM), In: Thermal simulation model for PCM energy storage and PCM wall construction
performance of the exterior envelopes of whole buildings XI international element. Ecostock conference; 2006.
conference. Clearwater Beach, FL; 2010. [160] Kuznik F, Virgone J, Johannes K. Development and validation of a new
[133] Tabares-Velasco PC, Christensen C. Bianchi MVA. Validation methodology to TRNSYS type for the simulation of external building walls containing PCM.
allow simulated peak reduction and energy performance analysis of Energy and Buildings 2010;42:1004–9.
residential building envelope with phase change materials, In: 2012 [161] Bontemps A, Ahmad M, Johanne s K, Sallée H. Experimental and modelling
ASHRAE annual conference. San Antonio, Texas; 2012. study of twin cells with latent heat storage walls. Energy and Buildings
[134] Hong T, Chou SK, Bong TY. Building simulation: an overview of develop- 2011;43:2456–61.
ments and information sources. Building and Environment 2000;35: [162] Thermal energy system specialists LLC -TESS Libraries—individual compo-
347–61. nent libraries. URL: /http://www.trnsys.com/tess-libraries/individual-com
[135] U.S.A. Department of Energy Building Energy Software Tools Directory. URL: ponents.phpS [accessed: October/23/2012].
/http://apps1.eere.energy.gov/buildings/tools_directory/S. [accessed: [163] University of strathclyde—energy systems research unit—ESP-r. URL:
October/23/2012]. /http://www.esru.strath.ac.uk/Programs/ESP-r.htmS. [accessed: October/
[136] Crawley DB, Hand JW, Kummert M, Griffith BT. Contrasting the capabilities 23/2012].
of building energy performance simulation programs. Building and Envir- [164] Heim D. Two solution methods of heat transfer with phase change within
onment 2008;43:661–73. whole building dynamic simulation,In: Proceedings of building simulation
[137] Barbour JP, Hittle DC. Modeling phase change materials with conduction 2005: 9th conference of international building performance simulation
transfer functions for passive solar applications. Journal of Solar Energy association. Montréal, Canada; 2005.
Engineering 2006;128:58–68. [165] Heim D, Clarke JA. Numerical modelling and thermal simulation of PCM–
[138] U.S.A. Department of Energy New Features in Version 2.0.0. URL: /http:// gypsum composites with ESP-r. Energy and Buildings 2004;36:795–805.
apps1.eere.energy.gov/buildings/energyplus/energyplus_archives.cfm#v4S. [166] Schossig P, Henning HM, Gschwander S, Haussmann T. Micro-encapsulated
[accessed: October/23/2012]. phase-change materials integrated into construction materials. Solar Energy
[139] Pedersen CO. Advance zone simulation in ENERGYPLUS: incorporation of Materials and Solar Cells 2005;89:297–306.
variable properties and phase change material (PCM) capability building [167] Fernandes NTA. Costa VAF. Use of phase-change materials as passive
simulation 2007,. China; 2007. elements for climatization purposes in summer: the Portuguese case.
[140] U.S.A. Department of Energy—EnergyPlus engineering reference. URL: International Journal of Green Energy 2009;6:302–11.
/http://apps1.eere.energy.gov/buildings/energyplus/pdfs/engineeringrefer [168] Almeida F, Zhang D, Fung AS, Leong WH. Investigation of multilayered
ence.pdfS. [accessed: October/23/2012]. phase-change-material modeling in ESP-R, In: International high perfor-
[141] Castell A, Medrano M, Castellón C, Cabeza LF. Analysis of the simulation mance buildings conference. Purdue; 2010.
models for the use of PCM in buildings. Effstock 2009 thermal energy [169] Sigalas G. Computational optimization of passive use of phase change
storage for efficiency and sustainability. Stockholm (Sweden),2009. materials in lightweight low-energy houses. Eindhoven: University of
[142] Muruganantham K. Application of phase change material in buildings: field Technology; 2011.
data vs. EnergyPlus simulation. Arizona State University; 2010. [170] AALBORG University—Danish Building Research Institute. URL: /http://
[143] Tardieu A, Behzadi S, Chen JJ, Farid MM. Computer simulation and experi- www.en.sbi.dk/publications/programs_models/bsim/about-bsimS.
mental measurements for an experimental PCM-impregnated office build- [accessed: October/23/2012].
ing, In: Proceedings of building simulation 2011: 12th conference of [171] Rose J, Lahme A, Christensen NU, Heiselberg P, Hansen M. Numerical
international building performance simulation association. Sydney; 2011. method for calculating latent heat storage in constructions containing
[144] Zhuang C, Deng A, Chen Y, Li S, Zhang H, Fan G. Validation of veracity on phase change material building simulation 2009, In: 11th international
simulating the indoor temperature in PCM light weight building by IBPSA conference. Glasgow, Scotland; 2009.
EnergyPlus. In: Li K, Fei M, Jia L, Irwin G, editors. Life system modeling [172] Corina S, Helmut EF, Frederick CW. Development of a simulation tool to
and intelligent computing. Berlin/ Heidelberg: Springer; 2010. p. 486–96. evaluate the performance of radiant cooling ceilings. Lawrence Berkeley
[145] Campbell KR. Phase change material as a thermal storage device for passive Laboratory Report, LBL-37300: University of California, Berkeley; 1995.
houses. Portland State University; 2011. [173] Kissock K. ESim building energy simulation software. Department of
[146] Chan ALS. Energy and environmental performance of building fac- ades mechanical and aerospace engineering. Dayton, OH: University of Dayton;
integrated with phase change material in subtropical Hong Kong. Energy 1997.
and Buildings 2011;43:2947–55. [174] Valentin energy software—PCMExpress. URL: /http://www.valentin.de/en/
[147] Kuznik F, Virgone J, Roux J-J. Energetic efficiency of room wall containing products/pcm/58/pcm-expressS. [accessed: October/23/2012].
PCM wallboard: a full-scale experimental investigation. Energy and Build- [175] Stetiu C. Radiant cooling in US office buildings: towards eliminating the
ings 2008;40:148–56. perception of climate-imposed barriers. University of California Berkeley;
[148] Shrestha S, Miller W, Stovall T, Desjarlais A, Childs K, Porter W, et al. 1997.
Modeling PCM-enhanced insulation system and benchmarking ENERGY- [176] Sowell EF, Haves P. Efficient solution strategies for building energy system
PLUS against controlled field data,In: Proceedings of building simulation simulation. Energy and Buildings 2001;33:309–17.
2011: 12th conference of international building performance simulation [177] Feustel HE, Stetiu C. Thermal performance of phase change wallboard
association. Sydney; 2011. for residential cooling application. Lawrence Berkeley Laboratory Report,
[149] Tabares-Velasco PC, Griffith B. Diagnostic test cases for verifying surface heat LBL-38320 University of California, Berkeley; 1997.
transfer algorithms and boundary conditions in building energy simulation [178] University of Dayton—Mechanical and Aerospace Engineering-ESim. URL:
programs. Journal of Building Performance Simulation 2011:1–18. /http://academic.udayton.edu/kissock/http/RESEARCH/EnergySoftware.
[150] Tabares-Velasco PC, Christensen C, Bianchi M. Verification and validation of htmS [accessed: October/23/2012].
EnergyPlus phase change material model for opaque wall assemblies. [179] Kelly Kissock J, Michael Hannig J, Whitney TI, Drake ML. Testing and
Building and Environment 2012;54:186–96. simulation of phase change wallboard for thermal storage in buildings. In:
[151] Bradley D, Kummert M. New evolutions in TRNSYS-A selection of version 16 Morehouse JM, Hogan RE, editors. International solar energy conference.
features, In: Proceedings of building simulation 2005: 9th conference of Albuquerque,: American Society of Mechanical Engineers; 1998.
international building performance simulation association. montréal, [180] Gatzka B, Valentin G. PCMexpress—planning and simulation programme for
Canada; 2005. the use of phase change materials (PCM) in buildings: demonstrating its use
[152] Ghoneim AA, Klein SA, Duffie JA. Analysis of collector-storage building walls in residential and office buildings in Ireland. International conference for
using phase-change materials. Solar Energy 1991;47:237–42. sustainable energy storage 2011. Belfast, UK; 2011.
[153] Ghoneim AA. Efficient collection and storage of solar energy. University Of [181] Colclough S, Griffiths P, Gschwander S Thermal energy storage and the
Alexandria; 1989. passive house standard: how PCM incorporated into wallboard can aid
[154] Stritih U, Novak P. Solar heat storage wall for building ventilation. Renew- thermal comfort, In: PLEA2009—The 26th conference on passive and low
able Energy 1996;8:268–71. energy architecture. Quebec City—Canada; 2009.